Table of Contents
Antonio e C. Paladini também presentes.
Contato com Walter: separar Data Publishing de Data Center por vezes é muito difícil... Tentativa de juntar esforços.
Dúvida: os voluntários são para ajudar no Data Publishing e Data Center, correto?? R: Sim.
Vinicius indisponível até meados de Fevereiro. Antonio demonstrou interesse, mas não avaliou em detalhes a parte técnica até meados de dezembro.
No "how to publish" do IVOA, só 2 toolkits apresentam conteúdo em Python: CVO e GAVO. Acionei o suporte de ambos e, como de costume no pessoal de VO, eles foram super prestativos. O CVO na verdade tem muita coisa em Java, nem tudo está livre (por agora?), enquanto que o GAVO tem tudo livre e excelente documentação. Nas palavras do Patrick Dowler (CVO):
Our approach differs somewhat from GAVO: opencadc provides a toolkit (libraries) from which you can assemble and customise VO services, while GAVO provides a more works-out-of-the-box framework that you can configure. We are moving to providing a better starting point through some example service projects and HOWTO documentation, but really our s/w is for shops that want to run highly scalable and robust services, customise the behaviour in some cases, and support a variety of authentication mechanisms and back-end systems.
Survey, sinergia com J-PAS e J-Plus.
BraVO publicação do S-Plus VO complaied. Paula está com data management do S-Plus. BraVO tem servidor para isso comprado pelo INCT-A no LAi. 1o. VO archive no Brasil.
Paula fazendo documentação. Ajuda bemvinda.
Pessoas vêem as páginas e se oferecem para ajudar.
2 voluntários potenciais: Nenhum é astrônomo. São da computação.
Vinicius, desenvolvedor de sistemas, querem participar das soluções de data publishing. Houve chamadas dos "Casos de Usuário", e ninguém foi capaz de resolver (comercialmente). Dados interoperados, públicos, ...
Para S-Plus, é necessário soluções comerciais.
Paula e Alex verão sobre a contribuição do Carlos Paladini (LAi).
Daniel assume o caso com Vinicius e Antonio para colaboração no data publishing.
Paulo Pessoal: definição para 6 meses.
1a (ou 2a) sextas-feiras do mês.
S-Data managements docs no GoogleDrive.
VO CEFCA (random notes)
implementation done from scratch in Python
- are there libraries?
JH: we used libraries (astropy, pywcs,...)
- are we considering shoaring our implementations?
JH: we should improve a bit the implementation
Both JH and Tamara never heard about VO before arriving here. None of them has ever attended an IVOA-interop meeting
JH: the development is "very specific" and oriented to our data
Hired
JH hired for db management and publishing (which ended up being VO)
Public and private access
Phase 3 data not public and not accessible via web-portal
Phase 2 will be private. Password protected data-access (certainly works with Topcat and Aladin)
protocols
Cone search, TAP, SIAP
SSAP still pending
Data models
Interops working on data-models on data-cubes ceFca could get involved at some point
Database
- Deisgned by JH and Tamara
It is postgresNG
- Why postgres?
JH: people here were using mySQL. We evaluated the possibilities but JH knew postgres which is known to manage big amount of data better than mySQL and with a bigger data model.
- Do IVOA db documents leave freedom of db implementation?
JH: totally free
Where does the system becomes "VO-compliant"?
- The db can be any. The Python layer receives VO-compliant requests, queries the db
and gives VO-compliant replies.
Registry
The service will be registered (possibly to the euro-vo registry with the help of SVO)
What is a VO?
VO was built on top of Internet standards, specially HTTP and XML, SOAP/WSDL or REST for description of the web services.
VO service term has a wide application, but usually means "anything that does something for you, implying a communication between two or more computers".
VO resource comprises web pages, database, storage element, interactive application - anything that is uniquely addressable through the Internet.
VO architecture
| USER LAYER | REGISTRY DATA ACCESS | (Finding) VO CORE (Getting) | PROTOCOLS | RESOURCE LAYER
Development according to the TCG (Technical Coordination Group).
TAP: Table Access Protocol, access to more complex search, as date, filters, etc.
SQL queries (very limited), lead to ADQL development (Astronomical Data Query Language), dealing with strings and XML formats.
VO do not dictate what deployers do internally, but rather simply to make them commit to standard interfaces.
VO Table is a exchange medium, not a primary storage format.
In VO exists a registry of registries.
IVOA needs highly technical standards to specify how services glue together. VOSI: VO Support Interfaces.
Web Services Basic Profile is a kind of guidebook for constructing web services that complies with both general web standards and IVOA standards. Universal Worker Service, JDL (Job Description Language).
Vo Space is the ability to share stored information through the VO standard.
VO philosophy is not to mandate any particular remote storage technology, but just to specify how to interface to such a remote storage system.
A participating system has to have a unique address following the vos://aaa.bbb.ccc/xxx form, to provide methods.
The initial stages of VO development relied on fully public data. However, much of astronomy involves data with proprietary access. This requires a standard definition of identity, methods for authenticating, authorize the used, etc.
VO uses the system of public key encryption, X509 certificates, TLS...
Rather that mega-application, it is more efficient if smaller apps specialize in particular jobs and communicate between each other. The ways are VO Tables and SAMP (Simple Application Message Protocol), a hub and spokes communication method.
Any SAMP-compatible application can start a Hub if there isn't one already running.
Depends on what you want to do.